Ultra-lightweight, high-fidelity text-to-speech for ComfyUI using Soprano TTS.
- Ultra-Fast: ~100-200x real-time generation
- High-Fidelity: 32kHz crystal-clear audio
- Lightweight: <1GB VRAM usage, 80M parameters
- All-in-One: Integrated saving and playback with autoplay
- ComfyUI Integration: Full memory management & interrupt support
- Easy Installation: Works with PyTorch 2.5.1 (no ComfyUI breakage!)
# 1. Navigate to ComfyUI and activate your Python environment
cd /path/to/ComfyUI
source venv/bin/activate # Linux/Mac
# venv\Scripts\activate # Windows
# 2. Install soprano-tts WITHOUT dependencies (critical!)
pip install soprano-tts --no-deps
# 3. Install required dependency
pip install unidecode
# 4. Restart ComfyUIThat's it! The node uses transformers backend (scipy and transformers already in ComfyUI).
- Add "Soprano TTS" node (audio/generation category)
- Enter text to synthesize
- Run workflow — audio plays automatically!
- (Optional) Connect audio output to other nodes for further processing
First run: Downloads model (~500MB) from HuggingFace, takes ~30 seconds, then cached.
| Parameter | Type | Default | Description |
|---|---|---|---|
| text | STRING | - | Text to convert to speech (2-15 sec sentences work best) |
| temperature | FLOAT | 0.3 | Lower = consistent, higher = varied (0.0-2.0) |
| top_p | FLOAT | 0.95 | Nucleus sampling threshold (0.0-1.0) |
| repetition_penalty | FLOAT | 1.2 | Reduces repetition (1.0-2.0) |
| seed | INT | -1 | Change to force regeneration (ComfyUI caching) |
| keep_model_loaded | BOOLEAN | False | Keep in VRAM for faster subsequent runs |
| Parameter | Type | Default | Description |
|---|---|---|---|
| filename_prefix | STRING | audio/soprano_tts | Path/prefix for saved files |
| format | ENUM | wav | Output format: wav, mp3, flac, ogg |
| bitrate | ENUM | 320k | Bitrate for mp3/ogg: 128k, 192k, 256k, 320k, VBR |
| save | BOOLEAN | True | Save to output folder (False = temp/preview only) |
| autoplay | BOOLEAN | True | Auto-play audio in the UI when generated |
- audio (AUDIO): 32kHz waveform compatible with ComfyUI audio nodes
- Built-in player: Audio plays directly in the node (with autoplay option)
- Auto-save: Files saved to
ComfyUI/output/audio/with counters
✅ DO:
- Use sentences 2-15 seconds long
- Convert numbers to words ("1+1" → "one plus one")
- Use proper grammar and contractions
- Regenerate (change seed) if result isn't perfect
❌ DON'T:
- Use very long or very short sentences
- Include special characters (%, #, @, etc.)
- Use all caps or improper grammar
- Expect identical output each time (uses sampling)
The Problem: Soprano TTS officially requires PyTorch 2.8+, but ComfyUI uses PyTorch 2.5.1. Installing normally upgrades PyTorch and breaks ComfyUI with errors like:
ImportError: undefined symbol: ncclMemFree
The Solution: Install with --no-deps to skip the problematic lmdeploy dependency. Soprano automatically falls back to the transformers backend, which works perfectly with PyTorch 2.5.1!
| Backend | Speed | PyTorch | Status |
|---|---|---|---|
| lmdeploy | ~2000x real-time | 2.8.0+ | ❌ Breaks ComfyUI |
| transformers | ~100-200x real-time | 2.5.1+ | ✅ Works! |
Even at 100-200x, generating 10 seconds of audio takes less than 1 second!
pip install soprano-tts --no-deps
pip install unidecodeYour PyTorch got upgraded accidentally. Fix it:
# Remove conflicting packages
pip uninstall -y lmdeploy torch triton
# Reinstall correct PyTorch for ComfyUI
pip install torch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 \
--extra-index-url https://download.pytorch.org/whl/cu121
pip install xformers==0.0.29.post1- Check internet connection and HuggingFace access
- Need ~500MB free disk space
- Ensure
huggingface-hubis installed:pip install huggingface-hub
- Lower
temperature(try 0.2) - Convert numbers/symbols to words
- Ensure sentences are 2-15 seconds
- Regenerate with different seed
- Restart ComfyUI completely
- Check console for loading errors
- Run test:
python test_compat.py
cd /path/to/ComfyUI/custom_nodes/ComfyUI-Soprano-TTS
python test_compat.pyShould output: ✅ ALL TESTS PASSED! Soprano TTS should work!
If tests fail, the script provides specific fix commands.
The node includes soprano_compat.py which:
- Checks PyTorch version (requires 2.5.0+)
- Patches
torch.compiler.disableif needed (usually not needed for 2.5.1) - Auto-applies on import
- Memory Management: Properly loads/unloads models, cleans CUDA cache
- Interrupt System: Respects user cancellation during generation
- Audio Format: Returns standard ComfyUI AUDIO (waveform + sample_rate)
- Model Caching: Downloads once, reuses forever
- Backend: transformers (HuggingFace)
- Speed: ~100-200x real-time
- Quality: 32kHz, perceptually indistinguishable from 44.1/48kHz
- VRAM: <1GB (80M parameters)
- Latency: First token in ~100ms, total generation <1s for 10s audio
- No streaming mode (transformers backend limitation)
- No voice cloning (Soprano limitation)
- No multilingual support yet (Soprano limitation)
- May occasionally mispronounce numbers/special characters
See workflows/example_workflow.json for a complete example.
This node wraps the excellent Soprano TTS project:
- Original Project: github.com/ekwek1/soprano
- Model: huggingface.co/ekwek/Soprano-80M
- Developer: ekwek1 (impressive work from a second-year undergrad!)
Soprano uses and/or is inspired by:
- OS: Linux or Windows
- GPU: CUDA-enabled (CPU support coming to Soprano)
- PyTorch: 2.5.0+ (tested with 2.5.1+cu121)
- VRAM: <1GB
- Disk: ~500MB for model cache
- pydub: For integrated audio saving/playback (
pip install pydub) - ffmpeg: Required by pydub for MP3/FLAC/OGG export
This ComfyUI node wrapper is licensed under Apache-2.0, matching the original Soprano project.
Contributions welcome! Please:
- Fork the repository
- Create a feature branch
- Submit a pull request
- Node Issues: Open an issue
- Soprano TTS Issues: Original project
- Full Soprano TTS integration
- PyTorch 2.5.1 compatibility via transformers backend
- ComfyUI memory management & interrupt support
- Automatic model downloading
- Comprehensive error handling and testing
Built with ❤️ for the ComfyUI community
Ultra-fast, high-fidelity speech synthesis in your workflows!